A bandgap reference generator is used to generate a constant bandgap reference voltage. This constant bandgap reference voltage may be temperature independent and can be used for different applications. For example, the constant bandgap reference voltage may be used in digital-to-analog converters (DACs), phase locked loops (PLLs), linear voltage regulators (VRs), DC-DC converters, RF circuits, etc.
Bandgap reference architecture is shown with reference to FIG. 1. Architecture 100 consists of a bandgap reference generator 101, a current source formed from p-type transistor MP1, and load 102 to receive the bandgap reference (bgref) voltage. Transistor p-type MP1, or a resistor, biased by Vbias can be used to provide supply current to load 102. One problem with architecture 100 is that because of process variations, current drawn by bandgap reference generator 101 and from node bgref (i.e., by load 102) can vary significantly. Thus the quiescent current of p-type transistor MP1 is designed with a lot of spare current providing capability.
This “spare” current capability becomes wasted current in most cases and can be two to three times waste compared to the minimum current needed by bandgap reference generator 101. Also, if there are current transients on node bgref, the response of bandgap reference generator 101 may be very slow, since bandgap reference generator 101 needs to go through the entire loop (not shown) of the bandgap reference generator 101, which may have limited bandwidth (typically, 20 MHz). Because of this low bandwidth, the power supply rejection ratio (PSRR) of architecture 100 at high frequencies is poor.